Human Occupation at Jinmium, Northern Australia: 116,000 Years Ago or Much Less?

Article excerpt

The rock-shelter of Jinmium in the Northern Territory of Australia hit the headlines a year-and-a-half ago when TL dates suggested human occupation might date from 116,000 years ago. Such dates were much earlier than any previously obtained for Australia, and thus suggested the continent was colonized at a very early stage in human dispersal around the Pacific. However, some TL dating is notoriously difficult to interpret, and here Nigel Spooner has re-assessed one of the later dates in the finmium sequence. His interpretation calls into doubt some of the earlier claims.

Introduction

Recently, Fullagar et al. (1996) have published in this journal a paper presenting evidence from thermoluminescence (TL) dating for human occupation of Australia at Jinmium, Northern Territory prior to 116,000 + or - 12,000 years. Additionally, artistic activity inferred from ochre is associated with TL dates of 75,000 to 116,000 years, and a rock-shelter section basal age of 176,000 [+ or -] 16,000 years was reported.

Here is presented an alternative interpretation of the luminescence data presented in Fullagar et al. (1996) for sample W1647, and as a result, a revised estimate for the age of that sample and consequently of the Jinmium rockshelter section. Discussion is focused on sample W1647 from the rock-shelter rather than on the deeper samples associated with the archaeologically more contentious samples, as W1647 is the only sample from the rock-shelter for which sufficient data is presented in Fullagar et al. (1996) to enable direct re-assessment of the age.

Sediment dating - resetting of TL by sunlight

The event measured by TL dating of unburnt sediment is the last exposure to sunlight of the mineral grains analysed. The quartz grains used in the dating at Jinmium have two principal TL peaks applicable to sediment dating: the rapidly bleached TL peak at 325 [degrees] C and the slowly bleached TL peak at 375 [degrees] C (these temperature are 'conventional' and relate to the temperature at which the TL peak is observed under the measurement conditions employed). Depositional contexts in which cumulative prolonged exposure to unattenuated sunlight is assured can be dated using either TL peak, as both will have been completely reset; this assumption is readily verified using the 'plateau' test. However, if incomplete bleaching is suspected, then the best measurement of the age of the sediment is found from the regrowth of the more light-sensitive 325 [degrees] C TL peak.

The 325 [degrees] C TL peak exhibits an extraordinary susceptibility to visible light, whereas the 375 [degrees] C TL peak is effectively unbleached by the same and requires exposure to ultra-violet (UV) light to bleach (Spooner et al. 1988; Spooner 1994). The difference in behaviour of these two TL signals is illustrated in FIGURE 1 by the response to UV-free 'white' light of the TL from quartz grains extracted from sediment approximately 20,000 years old, collected from Lake Woods, Northern Territory, Australia.

Resetting of TL at Jinmium

Although from a different locality, the Lake Woods sample has TL characteristics similar to those from Jinmium quartz, permitting comparison to be made with Fullagar et al. (1996): figure 21, which shows TL glow curves measured for sample number W1647 from excavation trench TL9 (in the rock-shelter) and for sample number W1752 from the sandsheet auger hole 5 TL14. FIGURE 2 shows an overlay of these two published glow curves; the TL glow-curve shapes are significantly different, with the rockshelter sample lacking a prominent 325 [degrees] C TL peak. It appears that since the last exposure to sunlight, the rapidly bleached 325 [degrees] C TL peak has barely regrown with respect to the 375 [degrees] C TL peak: therefore this last exposure must have taken place in the relatively recent past.

The effect of the 'missing' 325 [degrees] C TL peak on the TL plateau is shown in FIGURE 3, based on Fullagar et al. …